Indian Instrument Identification from Polyphonic Audio using KNN Classifier

Abstract

Classical Indian Music is considered as a very contrast and unique area of music across the globe. It has its own different melodies made up from Indian musical instruments which consists a wide range of variety in it. This paper basically deals with the Indian based musical instruments and presents an actual study on feature analysis for Indian musical instrument identification using machine learning techniques and to select and evaluate features extracted from a number of various feature scheme. All effort is employed to improve the different parts of the recognition system and our main work is linked to the features of the musical Instruments and the classification scheme includes K Nearest Neighbor (KNN). The obtained results using 7 Indian based musical instruments exhibit the good discrimination capabilities of the method.